SEARCH

SEARCH BY CITATION

Keywords:

  • Iron;
  • Pancreas;
  • Sickle Cell Disease;
  • Thalassemia;
  • Heart

Summary

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Iron endocrinopathy and cardiomyopathy are common in chronically transfused thalassaemia major patients, but relatively rare in chronically transfused patients with sickle cell disease. Since magnetic resonance imaging can demonstrate preclinical organ iron deposition, we hypothesized that pancreas and cardiac R2* would likewise be lower in sickle cell disease patients than thalassaemia major patients having comparable transfusional burdens. To test this hypothesis, we examined pancreatic and cardiac iron in a convenience sample of 100 chronically-transfused sickle cell disease and 131 thalassaemia major patients. Cardiac R2* (30 ± 9·2 vs. 73 ± 53 Hz, P < 0·0001) and pancreatic R2* (52 ± 62 vs. 253 ± 224 Hz, P < 0·0001) were significantly lower in sickle cell disease than thalassaemia major. Liver iron concentration was similar in both groups (14·9 ± 9·8 vs. 12·3 ± 8·4 mg/g dry weight, P = 0·101). The observed disparity in pancreatic and cardiac iron loading between sickle cell disease and thalassaemia major patients mirrors prior observations of organ toxicity in these patients. Greater cumulative transfusional iron exposure in thalassaemia major patients partially explains these observations but our data also suggest innate differences in labile iron handling between the two diseases.

Sickle cell disease (SCD) patients are increasingly being managed with chronic transfusions in order to prevent de novo or recurrent neurovascular complications (Mazumdar et al, 2007; Ware, 2007). SCD patients are generally transfused to maintain haemoglobin S levels of <30%, consequently lowering the risk of stroke (Vichinsky, 2001). This change in clinical care has led to similar transfusion requirements and comparable iron burdens in paediatric SCD and TM patients (Fung et al, 2008).

Despite comparable hepatic iron loading, TM and SCD patients have major differences in extra-hepatic iron toxicity (Vichinsky et al, 2005; Fung et al, 2006, 2007). In particular, diabetes and cardiomyopathy are relatively rare in chronically transfused SCD patients (Vichinsky et al, 2005; Fung et al, 2006, 2007). We previously demonstrated that cardiac iron loading was rare in SCD patients, consistent with their low prevalence of cardiomyopathy (Wood et al, 2004). To test the hypothesis that pancreas iron burden is similarly lower in SCD, we compared pancreatic, hepatic, and cardiac iron loading in 100 chronically transfused SCD patients and 131 patients with TM.

Methods

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

A convenience sample of clinical data was performed for all TM and chronically transfused SCD patients who had undergone clinically-indicated MRI examinations between 2004 and 2007; waiver of informed consent was granted by the Committee of Clinical Investigation at Children’s Hospital Los Angeles (CCI#07-00141). Prospective pancreatic MRI data was obtained from 22 control subjects. Informed consent was obtained for these patients (CCI#2000-076). 429 MRI studies (169 SCD and 260 TM) suitable for analysis were identified from 100 SCD and 131 TM patients. Data from the thalassaemia major patients and control subjects has been previously reported (Noetzli et al, 2009).

MRI acquisition and analysis methods for the liver, heart, and pancreas have been previously described (Wood et al, 2005a; Ghugre et al, 2006b). Cardiac and pancreas R2*, equal to 1000/T2*, was used for all statistical comparisons because R2* is proportional to tissue iron (Ghugre et al, 2006a; Wood et al, 2005b).

All MRI values were inversely weighted by the number of exams such that each patient contributed equally to correlation and group-wise analyses. Difference among the three study groups (TM, SCD, and controls) was assessed using Analysis of Variance with Dunnett’s post hoc correction. Weighted least-squares linear regression was used to assess the linear associations between liver iron concentration (LIC), pancreatic R2* and cardiac R2*. Weighted logistic regression with respect to age was performed to determine the time-varying prevalence of cardiac and pancreatic iron loading. All statistics were performed in jmp 5.1 (SAS, Cary, NC, USA).

Results

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Demographics of the study populations are summarized in Table I. TM patients were an average of 6·3 years older (P < 0·0001) but well matched with respect to LIC. Ferritin was more than 1000 ng/ml higher in SCD patients (= 0·02), but transferrin saturation was 21·6% lower (≤ 0·0001). Age of transfusion initiation was documented for 107/131 TM patients and 63/100 SCD patients. On average, TM patients started transfusions 5·5 years earlier than SCD patients (3·6 ± 5·9 vs. 9·1 ± 5·6 years, ≤ 0·0001), respectively. Accurate transfusional data was available only for 35 patients in each group. Median annual transfusion load was 199·3 ml/kg/year and 168·5 ml/kg/year for TM and SCD patients, respectively (P < 0·05 by Wilcoxin signed rank test). Pre transfusion hemoglobins were 0·4 g/dl lower in the SCD patients (= 0·03). Most of the patients were receiving monotherapy with deferasirox or deferoxamine; <5% were on some form of combined chelation therapy. Nearly 20% of SCD patients were on no iron chelation therapy compared with 5% of TM patients. Surgical splenectomy was more common in the TM patients (39/131 vs. 15/100, = 0·004). Splenectomy was not associated with increased extraheptic iron stores in SCD patients but splenectomized TM patients had increased pancreatic (360 ± 276 vs. 203 ± 189 Hz, = 0·015) and cardiac (92 ± 53 vs. 63 ± 51 Hz, = 0·005) R2* values. However, splenectomized TM patients were also older than non-splenectomized patients (23·1 ± 7·3 vs. 19·2 ± 7·6, = 0·005).

Table I.   Demographics of the study population.
 ControlsSCDTM
  1. LIC, liver iron concentration; MRI, magnetic resonance imaging; DFO, deferoxamine; DFX, deferasirox; DFP, deferiprone.

  2. *P < 0·05 with respect to controls.

  3. P < 0·05 with respect to SCD.

Age (years)29·7 ± 11·614·7 ± 4·6*21 ± 7·6†
Gender11F; 11M49F; 54M67F; 64M
Age onset of transfusion (years)Not applicable9·1 ± 5·6 (63/100 patients)3·6 ± 5·9† (107/131 patients)
Median transfusion Burden (ml/kg per year)Not applicable168·5199·3†
ChelatorNot applicable7 DFO; 52 DFX 2 DFX + DFO; 20 none 19 unknown28 DFO; 81 DFX; 2 DFX + DFO; 2 DFP; 1 DFP + DFX; 5 DFP + DFO; 5 none; 7 unknown
LIC by MRI (mg/g)1·0 ± 0·114·6 ± 9·8*12·3 ± 8·4*
Cardiac R2† (Hz)Not done30·4 ± 9·272·7 ± 52·9†
Pancreas R2† (Hz)21·8 ± 4·950·8 ± 61·6*253·3 ± 224·2*†
Pre-transfusion Haemoglobin (g/l)Not applicable94 ± 8·0 (66/100 patients)98 ± 8·0† (53/131 patients)
Ferritin (μg/l)Not done3943·2 ± 3463·6 (66/100 patients)2826·9 ± 2176·7† (109/131 patients)
Transferrin saturation (%)Not done58·4 ± 20·4 (53/100 patients)80·0 ± 20·2† (86/131 patients)
SplenectomizedNot applicable15/100 patients39/131 patients

Cardiac R2* was significantly lower in the SCD patients. Using the literature-accepted cardiac R2* cutoff of 50 Hz (T2* < 20 ms) to designate ‘detectable’ iron loading, there was a markedly lower prevalence of detectable cardiac iron in SCD patients than TM patients (1·5% vs. 37·7%).

Pancreatic R2* was markedly lower in the SCD patients compared to TM patients (52·0 ± 61·9 Hz vs. 252·5 ± 224·4 Hz, P < 0·0001) but was increased relative to control subjects (21·8 ± 4·9 Hz,  0·02). Using a 95% confidence interval for normal controls of 33·9 Hz, 25% of SCD patients and 72% of TM patients had abnormal pancreas R2* values (≤ 0·0001). Although our control group was not age-matched, pancreas R2* was independent of age in our control population.

Pancreas R2* was weakly associated with LIC for patients with SCD (r2 = 0·19, P < 0·0001) and TM (r2 = 0·13, P < 0·0001); results not shown. We previously demonstrated that pancreas R2* is strongly correlated with cardiac R2* in TM (r2 = 0·52, P < 0·0001; Noetzli et al, 2009). On ROC analysis, a pancreas R2* > 100 Hz was a strong predictor of concomitant cardiac iron loading (62% sensitivity, 97% specificity; Noetzli et al, 2009). In SCD, pancreas and cardiac R2* are also correlated (r2 = 0·26, P < 0·0001), but the relationship is driven by five examinations having increased cardiac iron (Fig 1). Four of the examinations were measurements in the same patient over a two and half year period; pancreas iron increased from 282 to 812 Hz while cardiac R2* increased only 20%, suggesting more rapid kinetics in the pancreas. The other patient with detectable cardiac iron also had severe pancreatic iron deposition (R2*c. 400 Hz); both patients had severe liver iron deposition (21 and 40 mg/g).

image

Figure 1.  Relationship between cardiac R2* and pancreatic R2* in SCD patients. Linear fit is shown (r2 = 0.26, P < 0.0001), although the relationship is driven by a few examinations with increased cardiac iron. Upper limit of pancreatic R2* (28·1 Hz) and cardiac R2* (50 Hz) are shown for reference (dashed lines). While many patients have increased pancreas iron and no detectable cardiac iron, the converse is not true; almost all patients with detectable cardiac iron have detectable pancreatic iron.

Download figure to PowerPoint

To investigate at what age significant pancreatic and cardiac iron develop, we performed logistic regression with respect to age for both populations (Fig 2) using pancreas R2* of 100 Hz (based on ROC analysis from TM patients) and cardiac R2* of 50 Hz (T2* of 20 ms) as cutoffs. The percent normal for pancreas iron (dashed line) and heart iron (solid line) are shown with SCD patients in black and TM patients in grey. The percentage of patients with normal cardiac and pancreatic iron decreases with increasing age in both populations; this decrease is greater in TM, as expected. Additionally, iron is detected earlier in the pancreas than in the heart for both TM and SCD patients, as noted by the leftward shift of the dashed line relative to the solid line. Using 50% prevalence as a reference, the time between detection of pancreas iron and cardiac iron is approximately 12 years in the TM population. Given the relatively lower loading of both organs in SCD patients and paucity of middle-aged patients, this difference could be estimated only through extrapolation, although the pattern was qualitatively similar. The extrapolated delay in pancreatic iron loading between SCD and TM patients, using 50% prevalence as a reference, was approximately 17 years.

image

Figure 2.  Logistic regression with respect to age for TM (grey) and SCD (black). The percentage of patients having pancreas R2* < 100 Hz (dashed line) and heart R2* < 50 Hz (solid line) are on the y-axis. Age is on the x-axis. Due to a lack of SCD patients older than 32 years, the data from ages 32–50 years is extrapolated (dotted line). The percentage of patients with insignificant cardiac and pancreatic iron decreases with increasing age in both populations; this decrease is greater in TM, as expected. Additionally, significant iron accumulates earlier in the pancreas than in the heart for both TM and SCD patients, as noted by the leftward shift of the solid line relative to the dashed line.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The purpose of this study was to determine whether the lower rate of diabetes and cardiac disease observed in chronically transfused patients with SCD was accompanied by lower pancreatic and cardiac iron burdens (Vichinsky et al, 2005; Fung et al, 2006). Our study clearly demonstrates that while pancreatic iron loading does occur in transfused SCD patients, it occurs later in life and to a lesser degree than for TM patients. Some of this difference can be attributed to the lower duration and intensity of transfusion therapy. In Multicentre Study Iron Overload (MCSIO), TM patients initiated transfusion therapy at a mean age of 4·3 ± 6·9 years, compared with 8·3 ±7·0 years for SCD patients transfused for stroke prophylaxis, P ≤ 0·0001 (Fung et al, 2006, 2007), comparable with the differences we observed (Table I).

In addition to beginning transfusion therapy at an earlier age, transfusion intensity is often higher in TM patients. In the MCSIO, annual transfusion volumes were 13·7% higher (= 0·07) in TM patients, compared with SCD patients, in the paediatric age group and 38% higher (P < 0·001) in TM adults (Fung et al, 2008). Even greater differences in transfusion rates were observed in the multicentre deferasirox trials sponsored by Novartis (59% greater for TM, P < 0·0001) (Cohen et al, 2008). In our patient population, mean and median annual transfusion volumes were only 9·5% and 18·2% higher (P < 0·05) for TM patients than SCD patients, respectively, consistent with the MCSIO data.

Using our estimates for differences in transfusion initiation (5·5 years), patient age (6·3 years) and transfusion intensity (85%), one would not expect more than a 13·9 year shift in the R2* logistic curves between TM and SCD patients, somewhat <17 year difference observed in the study population. The relative resistance of SCD patients to pancreatic and cardiac iron could not be explained by differences in chelator type or the total iron balance (as assessed by LIC).

Surgical splenectomy, which could potentially predispose to greater extrahepatic iron stores by eliminating a large physiologic iron depot, was more common in the thalassaemia major patients. In fact, splenectomy was associated with higher cardiac and pancreatic iron burdens, but not LIC or ferritin, in the TM patients. However, splenectomized patients were an average of 3·9 years older, confounding the comparison. Splenectomy was not associated with increased extrahepatic iron in SCD patients, although ours numbers were small. SCD patients also have a high prevalence of autosplenectomy and the effect of this process on extrahepatic iron stores is unknown (Brewer et al, 2009).

Pathologic loading of the heart and endocrine tissue occurs primarily or exclusively through circulating NTBI (Oudit et al, 2003; Glickstein et al, 2006). Transferrin saturations and NTBI levels are lower in transfused SCD patients compared with comparably iron loaded TM patients (Schein et al, 2008; Walter et al, 2008); we also noted lower transferrin saturations in our cohort. Hepcidin is the primary physiologic regulator of transferrin saturation and NTBI but no one has compared hepcidin levels in chronically transfused thalassaemia and sickle cell disease patients. Hepcidin is suppressed by increased erythropoetic drive, however our SCD patients were maintained at slightly lower pre-transfusion haemoglobin levels. Hepcidin is also upregulated by inflammation. Since sickle cell disease patients are chronically inflammed, upregulated hepcidin levels are a likely candidate to explain decreased transferrin saturations and extrahepatic iron deposition (Walter et al, 2008). However, prospective studies comparing hepcidin and cytokine levels in comparably transfused sickle cell and thalassaemia subjects will be necessary to test this hypothesis.

Regardless of the mechanism, cardiac iron deposition was only observed in 2/100 SCD patients. In fact, it is not clear whether it is cost-effective to routinely measure cardiac T2* values in SCD patients. Based on prior work in TM patients, we have been triaging patients into ‘low cardiac risk’ if their pancreas R2* values were <100 Hz. In the present SCD patient cohort, 80% of patients fall into this category. More importantly, all five examinations demonstrating cardiac iron exhibited pancreatic R2* > 100 Hz (481 ± 202·6 Hz, range: 282–821 Hz). Given the small number of SCD patients with cardiac iron, it is not possible to accurately assess algorithm effectiveness. In addition, pancreas R2* is more difficult to measure (interobserver 95% limits of agreement are 33%, at least double that for liver and heart iron measurements; Noetzli et al, 2009). Thus, this approach requires further clinical validation and does not supplant clinical acumen.

This retrospective study had several important limitations. Since the patients were referred for MRI from multiple centres and since the study was retrospective, clinical data regarding duration and intensity of transfusions was fairly limited, making it impossible to ‘match’ the SCD and TM populations for transfusional exposure. Our data represent a convenience sample; therefore, the different clinical indications for chronic transfusions as well as regional differences in clinical practice also confound comparisons between the two groups. Since many adult SCD patients discontinue chronic transfusion therapy, we were forced to extrapolate pancreatic and cardiac iron prevalence data for SCD patients >33 years of age, making estimates of time-delays between TM and SCD speculative. Although differences in hepcidin and NTBI are postulated as possible explanations for our observations, we did not make these measurements nor is the evidence linking NTBI to organ complications conclusive. Despite these limitations, pancreas R2* measurements in children are relatively straightforward to incorporate in clinical practice as markers of cardiac iron loading risk.

In summary, SCD patients have lower cardiac and pancreatic iron loading than chronically-transfused TM patients having similar liver iron levels. Much of this disparity can be explained by the larger transfusional burdens and durations observed in TM patients. However, innate differences in iron handling and elimination may also contribute to the large differences in extrahepatic iron loading observed in these patients.

Acknowledgements

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The authors would like to thank Jhansi Papudesi for her work on this project. In addition, we would like to thank Susan Carson, Anne Nord, Debbie Harris, Trish Peterson, Paola Pederzoli, Colleen McCarthy, Janelle Miller, Thomas Hofstra, and Susan Claster for their support of the MRI program.

References

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • Brewer, C.J., Coates, T.D. & Wood, J.C. (2009) Spleen R2 and R2* in iron-overloaded patients with sickle cell disease and thalassemia major. Journal of Magnetic Resonance Imaging, 29, 357364.
  • Cohen, A.R., Glimm, E. & Porter, J.B. (2008) Effect of transfusional iron intake on response to chelation therapy in beta-thalassemia major. Blood, 111, 583587.
  • Fung, E.B., Harmatz, P.R., Lee, P.D., Milet, M., Bellevue, R., Jeng, M.R., Kalinyak, K.A., Hudes, M., Bhatia, S. & Vichinsky, E.P. (2006) Increased prevalence of iron-overload associated endocrinopathy in thalassaemia versus sickle-cell disease. British Journal of Haematology, 135, 574582.
  • Fung, E.B., Harmatz, P., Milet, M., Ballas, S.K., De Castro, L., Hagar, W., Owen, W., Olivieri, N., Smith-Whitley, K., Darbari, D., Wang, W. & Vichinsky, E. (2007) Morbidity and mortality in chronically transfused subjects with thalassemia and sickle cell disease: a report from the multi-center study of iron overload. American Journal of Hematology, 84, 255265.
  • Fung, E.B., Harmatz, P.R., Milet, M., Balasa, V., Ballas, S.K., Casella, J.F., Hilliard, L., Kutlar, A., McClain, K.L., Olivieri, N.F., Porter, J.B. & Vichinsky, E.P. (2008) Disparity in the management of iron overload between patients with sickle cell disease and thalassemia who received transfusions. Transfusion, 48, 19711980.
  • Ghugre, N.R., Enriquez, C.M., Gonzalez, I., Nelson, M.D., Jr., Coates, T.D. & Wood, J.C. (2006a) MRI detects myocardial iron in the human heart. Magnetic Resonance in Medicine, 56. 681686.
  • Ghugre, N.R., Enriquez, C.M., Coates, T.D., Nelson, Jr, M.D. & Wood, J.C. (2006b) Improved R2* measurements in myocardial iron overload. Journal of Magnetic Resonance Imaging, 23, 916.
  • Glickstein, H., El, R.B., Link, G., Breuer, W., Konijn, A.M., Hershko, C., Nick, H. & Cabantchik, Z.I. (2006) Action of chelators in iron-loaded cardiac cells: accessibility to intracellular labile iron and functional consequences. Blood, 108, 31953203.
  • Mazumdar, M., Heeney, M.M., Sox, C.M. & Lieu, T.A. (2007) Preventing stroke among children with sickle cell anemia: an analysis of strategies that involve transcranial Doppler testing and chronic transfusion. Pediatrics, 120, e1107e1116.
  • Noetzli, L.J., Papudesi, J., Coates, T.D. & Wood, J.C. (2009) Pancreatic iron loading predicts cardiac iron loading in thalassemia major. Blood, 114, 40214026.
  • Oudit, G.Y., Sun, H., Trivieri, M.G., Koch, S.E., Dawood, F., Ackerley, C., Yazdanpanah, M., Wilson, G.J., Schwartz, A., Liu, P.P. & Backx, P.H. (2003) L-type Ca(2+) channels provide a major pathway for iron entry into cardiomyocytes in iron-overload cardiomyopathy. Nature Medicine, 9, 11871194.
  • Schein, A., Enriquez, C., Coates, T.D. & Wood, J.C. (2008) Magnetic resonance detection of kidney iron deposition in sickle cell disease: a marker of chronic hemolysis. Journal of Magnetic Resonance Imaging, 28, 698704.
  • Vichinsky, E.P. (2001) Current issues with blood transfusions in sickle cell disease. Seminars in Hematology, 38, 1422.
  • Vichinsky, E., Butensky, E., Fung, E., Hudes, M., Theil, E., Ferrell, L., Williams, R., Louie, L., Lee, P.D. & Harmatz, P. (2005) Comparison of organ dysfunction in transfused patients with SCD or beta thalassemia. American Journal of Hematology, 80, 7074.
  • Walter, P.B., Macklin, E.A., Porter, J., Evans, P., Kwiatkowski, J.L., Neufeld, E.J., Coates, T., Giardina, P.J., Vichinsky, E., Olivieri, N., Alberti, D., Holland, J. & Harmatz, P. (2008) Inflammation and oxidant-stress in beta-thalassemia patients treated with iron chelators deferasirox (ICL670) or deferoxamine: an ancillary study of the Novartis CICL670A0107 trial. Haematologica, 93, 817825.
  • Ware, R. (2007) Principles and indications of chronic transfusion therapy for children with sickle cell disease. Clinical advances in hematology and oncology, 5, 686688.
  • Wood, J.C., Tyszka, J.M., Ghugre, N., Carson, S., Nelson, M.D. & Coates, T.D. (2004) Myocardial iron loading in transfusion-dependent thalassemia and sickle-cell disease. Blood, 103, 19341936.
  • Wood, J.C., Enriquez, C., Ghugre, N., Tyzka, J.M., Carson, S., Nelson, M.D. & Coates, T.D. (2005a) MRI R2 and R2* mapping accurately estimates hepatic iron concentration in transfusion-dependent thalassemia and sickle cell disease patients. Blood, 106, 14601465.
  • Wood, J.C., Otto-Duessel, M., Aguilar, M., Nick, H., Nelson, M.D., Coates, T.D., Pollack, H. & Moats, R. (2005b) Cardiac iron determines cardiac T2*, T2, and T1 in the gerbil model of iron cardiomyopathy. Circulation, 112, 535543.